Alibaba DAMO Academy has open-sourced Damo Radar, a vision-language model designed to read contrast-enhanced computed tomography scans across 18 abdominal organs and identify nearly 150 conditions, including cancers and other abnormalities1.
The model analyzes contrast-enhanced CT scans to detect a broad range of diseases and abnormalities across the abdominal region. By open-sourcing the model, Alibaba Group Holding's research arm is making the diagnostic tool available for external use, a move described as part of the company's growing medical AI efforts.
Damo Radar is built as a vision-language model, a category that pairs image interpretation with natural-language reasoning. Its coverage spans 18 abdominal organs, and its condition library reaches nearly 150 distinct diagnoses, encompassing malignant findings alongside other pathologies. The model's input format is contrast-enhanced CT, a widely used but interpretation-intensive imaging modality in clinical radiology.
The release extends Alibaba Group Holding's AI activity beyond the cloud and enterprise tooling that has dominated its recent public profile. Alibaba Cloud Intelligence recently contributed its MuleRun and WonderClip platforms to Malaysia's AI untuk Rakyat youth program[2], and an Alibaba Group Holding model featured in autonomous-agent safety research published earlier this month[1].
ANALYSIS Open-sourcing a medical imaging model of this diagnostic breadth puts Alibaba DAMO Academy's work into direct comparison with other open medical-AI efforts, where the practical barrier is typically regulatory validation rather than model availability.